The US agency has warned employers to check ‘bossware’ tools for discrimination

The head of the US agency charged with enforcing civil rights in the workplace says artificial intelligence-driven “bosware” devices that closely monitor the location, keystrokes and productivity of workers could run afoul of discrimination laws.

Charlotte Burroughs, chairwoman of the Equal Employment Opportunity Commission, told The Associated Press that the agency is trying to educate employers and technology providers about the use of these intelligence tools, as well as AI tools that streamline job evaluations.

If you’re not careful about draconian scheduling algorithms that penalize pregnant women or Muslims for taking breaks or taking time to pray, or allowing flawed software to filter out graduates of women’s or historically black colleges — you can’t blame AI. The EEOC is calling.

“I’m not shy about using our enforcement authority when necessary,” Burroughs said. “We want to work with employers, but they can’t be exempt from civil rights laws because they discriminate in high-tech ways.”

A federal agency released its latest guidance Thursday on using automated systems to make employment decisions such as who to hire or promote. It explains how to interpret a key provision known as Title VII of the Civil Rights Act of 1964, which prohibits employment discrimination based on race, color, national origin, religion or sex, which includes discrimination against gay, lesbian and transgender workers.

Burroughs said one important example involves the widely used real screens and their ability to generate biased results based on biased data.

“What happens is you have an algorithm that looks for patterns that reflect patterns it already knows,” she said. “Training is based on data from existing employees. And if you currently have a diverse workforce, you could end up inadvertently firing people who don’t look like your current employees.”

Amazon, for example, abandoned its own recruiting tool to hire top talent after finding that men were preferred for technical roles — in part because it compared job candidates to the company’s male-dominated tech workforce.

Other agencies, including the Justice Department, have issued similar warnings over the past year about how some AI tools could discriminate against people with disabilities and violate the Americans with Disabilities Act.

In some cases, the EEOC has taken action. In March, tech job search site Dice.com agreed with the agency to drop an investigation into allegations that it allowed job posters for immigrants seeking work visas to exclude U.S.-citizen workers. To resolve the issue, parent company DHI Group agreed to rewrite its programs to “scratch” for discriminatory language such as “H-1Bs Only,” which refers to a type of work visa.

Much of the EEOC’s work involves investigating complaints filed by employees who believe they have been discriminated against. Although it’s hard to know if a biased hiring tool has caused job applicants to be denied a job, Burrows said there’s “generally more awareness” among employees about the tools they use to track their productivity.

Those devices can range from radio frequency devices to monitor nurses, monitor keystrokes or computer mouse clicks as many office workers begin working from home during the pandemic, to monitor warehouse workers and delivery drivers on a minute-by-minute basis. Some may violate civil rights laws depending on how they are used.

Burroughs noted that the National Labor Relations Board is looking at such AI tools. The NLRB issued a warning last year that overly intrusive surveillance and management tools could impair employees’ right to communicate with one another about union activity or dangerous situations.

“I think the best approach out there — I’m not saying don’t use it, it’s not illegal — but to really think about what employers want to measure and maybe measure that directly,” Burroughs said. “If you’re trying to see if the job is getting done, maybe check if the job is getting done.”